Context : Direct detection of a planet around a star and its characterization for identification of bio-tracers in the mid-IR , requires a nulling interferometer .
Such an instrument must be efficient in a large wavelength domain in order to have the capability to detect simultaneously the infrared spectral features of several bio-tracers : CO _ { 2 } , O _ { 3 } and H _ { 2 } O .

Aims : A broad wavelength range can be effective , provided that an achromatic phase shift of \pi can be implemented , with an accuracy good enough for achieving a deep nulling at all considered wavelengths .
A new concept for designing such an achromatic phase shifter is presented here .
The major interest of this solution is that it allows a simple design , with essentially one device per beam .

Methods : The heart of the system consists in two cellular mirrors where each cell has a thickness which introduces , for a given central wavelength , a phase shift of ( 2 k + 1 ) \pi or of 2 k \pi on the fraction of the wave it reflects .
Each mirror is put in one of the collimated beams of the interferometer .
Because of the odd/even distribution , a destructive interference is produced on axis for the central wavelength when recombining the two beams .
Now , if the number of cells of a given thickness follows a rather simple law , based on the Pascal ’ s triangle , then we show that the nulling is also efficient for a wavelength not too far from the central wavelength .

Results : The effect of achromatization is the more efficient the larger the number of cells is .
For instance , with two mirrors of 64 \times 64 cells , where the cells phase shift ranges between -6 \pi and +6 \pi , one reaches a nulling of 10 ^ { -6 } on a wavelength range [ 0.6 \lambda _ { 0 } , 1.25 \lambda _ { 0 } ] , i.e .
on more than one complete octave .
This is why we claim that this device produces a quasi-achromatic phase shift : especially , it could satisfy the specifications of space mission as DARWIN .
In a second step , we study the optimum way to distribute the cells in the plane of the pupil .
The most important criterion is the isolation of the planet image from the residual image of the star .
Several algorithms are presented , one being especially efficient and we present the nulling performances of those various configurations .

Conclusions :